In late November when Amazon announced that it was launching a machine learning service that will aim to mine data from electronic health records (EHRs), company officials, in a blog post, noted that the software, Amazon Comprehend Medical, is “a HIPAA-eligible machine learning service that allows developers to process unstructured medical text and identify information such as patient diagnosis, treatments, dosages, symptoms and signs, and more.”
Two of Amazon’s specialists—Matt Wood, Ph.D., a machine learning expert, and Taha Kass-Hout, M.D., a former FDA chief health informatics officer—noted that a core issue in healthcare and health IT today is that a large amount of critical data is stored as unstructured medical text, such as medical notes, prescriptions, audio interview transcripts, and pathology and radiology reports. “This means that being able to identify this information can be a manual and time-consuming process, which either requires data entry by high skilled medical experts, or teams of developers writing custom code and rules to try and extract the information automatically,” they outlined, adding that Comprehend Medical will instead aim to specifically allow developers “to identify the key common types of medical information automatically, with high accuracy, and without the need for large numbers of custom rules.”
Following the announcement, industry observers were quick to react to Amazon’s latest aggressive push into healthcare, with one consultant for Impact Advisors, Liam Bouchier, a principal with the firm, noting that this latest initiative is simply another example of Amazon “doing what it does best”—working to analyze large data sets as a means to gain meaningful insights into the consumer through a variety of different ways. “This concept does make some in the healthcare industry uncomfortable,” he added.
Another healthcare consultant—Michael Abrams, managing partner at the St. Louis-based Numerof & Associates—agrees that the initiative is exciting at the surface, and believes it speaks to a much larger issue in the health IT market: responding to a significant need to do something constructive with the voluminous amount of data that stakeholders already own.
Taking a step back, Abrams, in a recent interview, explains that one of the core issues that has arisen from healthcare’s digital shift is that having access to digitized data is different from being able to use that data to streamline decisions or improve quality. “Because hospitals rushed to digitization in response to extrinsic [financial] incentives [as part of ARRA/HITECH], they wanted to qualify for federal subsidies, but hadn’t really developed the internal capability to use the data. So, for the most part they are still very early on that learning curve, and the notion of aggregating data, data manipulation, and the insights you can draw from aggregating data, is still a new concept, Abrams says. He adds, “Now that [organizations] have these [EHR] systems and are saddled with substantial upkeep, it’s important to find a way to make it pay off.”
He contends that much of the healthcare delivery community is obsessed with “big data,” but the truth is that “they have hardly scratched the surface using the data they already own.” Abrams asserts, “Many of them don’t understand their own internal operations and don’t understand the cost of doing business— what it costs them to do a knee replacement, for example. All they know is at the end of the year, if they are profitable, it’s all good. Every corner convenience store has SKUs on everything they sell, and they know what the profitably is on every item. How many hospitals can say that? Not many,” he attests.
But according to what Amazon has said regarding its machine learning capabilities, this new software can re-digitize patient records and other clinical notes, analyze them, and pull out key data points, Abrams explains. “This is a very strategic move for Amazon because it already has advanced capabilities in natural language processing. Amazon Web Services has been selling this kind of text analysis software to markets outside of medicine for some time. They have a significant advantage,” he says.
A Vendor Community Satisfied with the Status Quo
An updated brief from the Office of the National Coordinator for Health IT (ONC) last summer revealed that health IT giants Epic and Cerner have continued to maintain the largest EHR market share (22 percent and 24 percent, respectively), based on the proportion of hospitals that reported using the developer's certified products.
These results, though unsurprising, signify to Abrams an oligopoly in that Epic and Cerner account for about half of the hospital market share, with the top-five vendors in this space accounting for roughly 85 percent of the market share. “When you have an oligopoly, you have a lot of large, slow-moving entities,” says Abrams, who also believes that as consolidation increases in the provider market, leading to fewer larger players among hospital systems, it makes it “almost incumbent upon vendors to also become fewer and larger, in order to have a balance of power.”
What’s more, Abrams attests that the EHR vendors with the largest market shares “regard their propriety code and the closed nature of their systems as a defense against encroachment by other, perhaps hungrier players who don’t have that market share. The status quo of a lack of interoperability suits the dominant players,” he says.
Can New Entrants Be Slowed Down?
With all this in mind, while health IT vendors have traditionally been content with incremental change, Abrams strongly believes that outsiders such as Amazon, Apple, and Google “bring the technology and commercial savviness to dramatically shake up what’s been generally a very complacent industry.” He adds, “Players like Amazon have their eye on the next big thing.”
Indeed, the traditional players and the healthcare delivery organizations are, on at least at some level, “terrorized about what these technology companies are going to come up with next,” he says. Consider that in just the last year alone:
- Amazon, Berkshire Hathaway, and JPMorgan Chase & Co announced they were teaming up on an initiative to improve satisfaction and reduce costs for their companies’ employees.
- Amazon said it would be part of another endeavor related to healthcare—to remove interoperability barriers and to make progress on adoption of health data standards. For this project, Amazon is teaming up with Microsoft, Google, IBM, and others to jointly commit to support healthcare interoperability by advancing healthcare standards such as HL7 (Health Level Seven International), FHIR (Fast Healthcare Interoperability Resources), and the Argonaut Project.
- Apple, of course, also had a big announcement in early 2018: that it would be testing its new Health Records feature out with 12 hospitals, inclusive of some of the most prominent healthcare institutions in the U.S. Since that time, more than 100 new organizations have joined the project, according to Apple. The idea behind the feature is that consumers could see their medical records right on their iPhones.
As such, prior to the past year when non-traditional players began moving more and more into healthcare, providers were “reasonably comfortable with moving as slowly as they could because they thought they had a good fix on the level of pressure and the pace of change that they might expect from the government,” says Abrams.
Adding to this thought, he notes that providers—with the help of industry trade associations—could negotiate with the government on the speed at which certain regulations might come down the pike, depending on how ready stakeholders are. But, Abrams adds, “Nobody can say ‘slow down’ to Amazon; when they want to do something, they are going to do it, and if they have a better solution, it could very well revolutionize the industry.”